For exercise prescription purposes, it is often assumed that % heart rate reserve (%HRR) provides equivalent intensities to %(V) over dotO(2max). However, a recent study from this laboratory demonstrated that during cycling exercise %HRR is not equivalent to %(V) over dotO(2max) but is instead equivalent to a percentage of the difference between resting and maximal (V) over dotO(2), i.e., %(V) over dotO(2) reserve (%(V) over dotO(2R)). The current study examined these relationships during treadmill exercise. Fifty adults performed Bruce protocol treadmill tests to exhaustion. For each subject, data obtained at rest, at the end of each stage, and at maximum were used to determine linear regressions of %HRR versus %(V) over dotO(2max) and of %HRR versus %(V) over dotO(2R). For %HRR versus %(V) over dotO(2max), the mean intercept and slept were -6.1 +/- 0.7 and 1.10 +/- 0.01, respectively, with a mean r of 0.900 +/- 0.002. For %HRR versus %(V) over dotO(2R), the mean intercept and slope were 1.5 +/- 0.6 and 1.03 +/- 0.01, respectively, with a mean r of 0.990 +/- 0.002, Both regressions differed statistically from the line of identity (i,e., intercept of 0 and slope of 1). However, the regression of %HRR versus %(V) over dotO(2R) was significantly closer (P < 0.001) to the line of identity than was the regression of %HRR versus %(V) over dotO(2max). We conclude that %HRR should be considered as an indicator of %(V) over dot(2R), not %(V) over dotO(2max), when prescribing treadmill exercise, as was previously concluded for cycling exercise.